SmartLearn AI

SmartLearn AI is a next-generation cognitive architecture and educational assistant. Built from the ground up to deliver uncompromising speed, reliability, and intelligence, the platform seamlessly integrates a sophisticated Retrieval-Augmented Generation (RAG) pipeline with real-time multi-modal AI capabilities.

Currently Live at: smartlearn.work

Enterprise-Grade Architecture

SmartLearn AI transcends traditional chatbot interfaces by offering a suite of industry-level tools designed for deep analytical research and interactive learning.

Core Capabilities

• Zero-Latency Live Code Execution: An integrated browser-based IDE powered by Sandpack. Users can write, execute, edit, and hot-reload React and JavaScript applications directly inside the chat interface without external dependencies.
• Autonomous Web Browsing (Playwright): SmartLearn launches a headless Chromium browser in the background to navigate URLs, wait for page renders, scrape content, and snap live viewport screenshots directly into the chat stream.
• Zero-Retention Private Mode: A strict, SOC2-compliant hardware-level privacy feature that bypasses the database completely. Conversations live exclusively in RAM and are permanently destroyed upon closing the session.
• Multimodal Vision Engine: Features a highly optimized client-side compression algorithm that processes high-resolution images instantly, routing complex visual data directly to specialized native models (e.g., Gemini 2.5 Flash) while completely avoiding backend token bloat.
• Advanced Fallback Architecture: Engineered for 100% uptime. The proprietary router dynamically cascades queries across elite foundational models (including Groq LLaMA 3.3, Google Gemini, and OpenRouter variants).
• Interactive 3D Flashcards: Features a mathematically robust parsing engine that sanitizes LLM-hallucinated markdown blocks, rendering study materials into pure CSS 3D perspective glassmorphic flashcards.
• Visual Knowledge Architecture: Automatically extracts complex relationships from uploaded documents and plots them into a living, interactive node graph utilizing React Flow.
• Advanced Voice Mode Engine: A hyper-realistic, hands-free conversational interface featuring a full-screen dynamic glowing orb. Built with auto-listen mechanics, phase-state orchestration (Listening, Processing, Speaking), and aggressive Whisper audio filters.
• Enterprise Email Infrastructure (Resend): A completely overhauled automated email engine featuring official SmartLearn Red branding, dynamic logo injection, and glassmorphic OTP verification boxes.
• Built-in Bug Reporting: A seamless, ChatGPT-style bug reporting modal directly in the UI that instantly structures user feedback and directly emails the admin system using our Resend integration.
• Performance & Traffic Analytics: Fully integrated with Vercel Web Analytics and Speed Insights to track global latency and user engagement in real-time.

Technical Stack

The infrastructure is meticulously separated into a high-performance Python backend and a lightning-fast React frontend.

Frontend

• React 19 (Vite)
• TailwindCSS (Utility-first styling with custom glassmorphism)
• Framer Motion (Fluid 60fps animations and state transitions)
• Vercel Web Analytics & Speed Insights

Backend

• FastAPI (High-throughput asynchronous routing)
• SQLAlchemy & Neon PostgreSQL (Persistent, encrypted data storage)
• FAISS & Sentence Transformers (In-memory semantic vector search)
• Upstash Redis (Global edge caching for sub-millisecond retrieval)
• Resend (Transactional emails and automated bug reports)

Development & Deployment

SmartLearn AI is designed for seamless local development using Docker Compose and optimized for production deployment on Vercel (Frontend) and Railway (Backend).

1. Local Setup via Docker (Recommended)

You can spin up the entire frontend and backend ecosystem instantly:

git clone https://github.com/Nav33dCodes/smartLearn-AI.git
cd smartLearn-AI

# Create your .env files in frontend and backend (see templates below)

# Boot the entire stack
docker-compose up -d --build

• Frontend will be available at http://localhost:5173
• Backend API will be available at http://localhost:8000

2. Manual Local Setup

Backend Environment:

cd smartlearn-backend
pip install -r requirements.txt
uvicorn main:app --reload

Frontend Environment:

cd smartlearn-frontend
npm install
npm run dev

3. Production Deployment (Vercel & Railway)

Frontend (Vercel):

1. Import the repository into Vercel.
2. Set the Root Directory to smartlearn-frontend.
3. Add your VITE_API_URL environment variable pointing to your Railway backend.

Backend (Railway):

1. Import the repository into Railway.
2. Set the Root Directory to /smartlearn-backend so Railway detects the production-ready Dockerfile.
3. Add your backend environment variables (including Database, Redis, and API keys). Railway will automatically map the $PORT.

Environment Variables Template

Create .env inside smartlearn-backend/:

Click to expand Backend Environment Template

# CORS & Routing
ALLOWED_ORIGINS="http://localhost:5173,https://smartlearn.work"
FRONTEND_URL="https://smartlearn.work"

# Security & DB
JWT_SECRET="your_secure_random_string_here"
DATABASE_URL="postgresql://user:password@host:port/dbname"
REDIS_URL="rediss://default:password@host:port"

# Email Configuration (Resend)
SMTP_EMAIL="noreply@smartlearn.work"
RESEND_API_KEY="re_your_api_key"

# AI Model Integrations
GROQ_API_KEY="gsk_..."
GEMINI_API_KEY="AQ..."
OPENROUTER_API_KEY="sk-or-v1-..."
TAVILY_API_KEY="tvly-..."
YOUTUBE_API_KEY="AIza..."

Create .env inside smartlearn-frontend/:

VITE_API_URL=https://your-railway-url.up.railway.app

Leadership & Engineering

• Sanan Malik – CEO & Visionary
• Naveed Ahmed – Lead Architect & Developer

License

Distributed under the MIT License. See LICENSE for more information.